Iterative component mode synthesis using a priori and a posteriori criteria

An iterative algorithm of component mode synthesis for a rational reduced-order modeling is proposed. The present algorithm consists of three steps: Initial mode selection with a priori criterion, model reduction using the selected modes, and a posteriori error evaluation. The algorithm is based on the Craig-Bampton method, which is one of the most widely used techniques. Specifically, the present algorithm uses, in a judicious manner, both an a priori mode selection criterion employing the moment-matching strategy and an a posteriori error estimation technique in evaluating the relative eigenvalue errors. If desired, an optional step of boundary interface reduction offers an additional possibility of obtaining a more compact model. The proposed iterative algorithm is recommended to researchers as well as commercialized software users in developing linear reduced-order structural dynamics models. Numerical examples are presented for its practical usage and performance.